The use of peptides as vaccines for the prevention or treatment of cancer has received considerable attention in recent years as further insight is gained into the steps required to elicit an effective immune response. Even though a greater understanding of these steps has been gained, the success of peptide-based vaccines in the clinic has been very limited. This lack of success in treating cancer is probably due to multiple factors, including the target proteins selected, the nature of the peptides chosen from these proteins and the local environment of the tumor. The selection of proteins as targets for the treatment of cancer is complicated by the fact that many or most tumor antigens that have been identified to date are non-mutated self-antigens, and that the immune response to these self-antigens is restricted by T-cell tolerance. T-cell tolerance can be mediated either centrally or peripherally, and results in the loss of high-avidity cytotoxic T-lymphocytes (CTL) from the immune repertoire that would have the potential to mediate the killing of tumor cells. Consequently, one challenge for the development of effective cancer vaccines is to identify epitopes in tumor antigens to which the immune system has not been tolerized or where tolerance can be overcome.
Unlike antibodies that recognize intact soluble or cell-bound proteins, T-cells recognize fragments of proteins (peptides) that are generated by proteolytic degradation. For CTLs, the peptides generally are comprised of 8 to 11 amino acid residues and are only recognized by cytotoxic T-cells when the peptides are bound to the class I major histocompatibility complex (MHCI). MHCIs are expressed by most cell types and play a critical role in the determination by the immune system of whether a cell is “self” or “non-self” (Whiteside, T. L. and Herberman, R. B., 1995, Curr. Opin. Immunol. 7, 704-711). The steps involved in processing an antigenic protein to generate peptides that can bind to and be presented by MHCIs have been extensively studied. These steps include: (i) internalization of the protein by an antigen-presenting cell (APC); (ii) degradation of the protein into peptides by the APC proteasome; (iii) translocation of the peptides in the endoplasmic reticulum by TAP transporters; and (iv) association of the peptides with the two chains of an MHCI to form a stable MHCI-peptide complex, which is then exported to the cell surface.
Several factors determine which peptides from an antigenic protein are bound to the MHCI and displayed on the cell surface. The ability to be recognized and cleaved by the proteasome, the stability of the peptide in the cytosol, and the efficiency of transport by the TAP transporters are all critical parameters; however, the most important determinant appears to be the ability of the peptide to bind to the MHCI (Chen et al., 1994, J. Exp. Med. 180, 1471-1483). The topology of the class I binding site determines which peptides can bind to the MHCI, and the binding site topology differs among the different class I molecules. Within the set of peptides able to bind to a given class I molecule, there is typically a range of binding affinities such that those with the highest affinity compete the best for display at the cell surface. Peptides preferentially selected by these processes for presentation at the cell surface are referred to as immunodominant epitopes. In general, peptides that bind with high affinity to the MHCI tend to elicit stronger CTL responses than peptides binding with lower affinity (Sette et al., 1994, J. Immunol. 153, 5586-5592). However, it is these higher affinity, immunodominant epitopes derived from self-proteins (such as cancer antigens) to which T-cells have been exposed and tolerized.
Several investigators have developed approaches for identifying CTL epitopes having low MHCI binding affinity, and increasing the binding affinity of these epitopes by making modifications to the native peptide sequence (Parker et al., 1992, J. Immunol. 149, 3580-3587; Tourdot et al., 1997, J. Immunol. 159, 2391-2398; Dionne et al., 2004, Cancer Immunol. Immunother. 53, 307-314). These and other investigators have shown that such modified peptides can be more immunogenic and capable of eliciting enhanced CTL responses in both in vitro and in vivo assays.
In addition to binding affinity, data from other laboratories have shown that the stability of the MHCI-peptide complex is important in determining the immunogenicity of CTL epitopes. Using two CTL epitopes from the Epstein-Barr virus nuclear antigen, it was demonstrated that the epitope forming the more stable MHCI-peptide complex elicited a stronger CTL response when lymphocytes isolated from EBV-seropositive donors were stimulated with a virus-transformed autologous lymphoblastoid cell line (Levitsky et al., 1996, J. Exp. Med. 183, 915-926). From studies examining the binding affinities and dissociation rates of a group of HIV-1 MHCI-binding peptides, it was concluded that the immunogenicity of the HIV-1 derived peptides might be predicted more accurately by the dissociation rate than by the binding affinity of the peptide to MHCI (van der Burg et al., 1996, J. Immunol. 156, 3308-3314). Furthermore, in studies examining multiple modified analogs of a murine p53 epitope, it was observed that modifications enhancing the stability of the MHCI-peptide complex resulted in significantly more potent immunogens (Baratin et al., 2002, J. Peptide Sci. 8, 327-334).
Although modification of CTL epitopes can be effective in increasing the binding affinity of the epitope to the MHCI complex and enhancing the immunogenicity of the epitope, various investigators have shown that these modifications can alter the specificity of the elicited immune response when compared with that of the unmodified “parent” peptide. Substitutions at amino acid “anchor residues” have been shown to alter both the conformation of the MHCI-peptide complex and the conformation of the contacts with the T-cell receptor (Sharma et al., 2001, J. Biol. Chem. 276, 21443-21449; Denkberg et al. 2002, J. Immunol. 169, 4399-4407). Alterations in the sequences of CTL epitopes have also been shown to change the specificity of the immune response elicited in vitro and in patients treated with the modified epitopes (Clay et al., 1999, J. Immunol. 162, 1749-1755; Yang et al., 2002, J. Immunol. 169, 531-539; Dionne et al., 2003, Cancer Immunol. Immunother. 52, 199-206; Okazaki et al., 2003, J. Immunol. 171, 2548-2555). These results indicate that when making modifications to amino acid sequences of low affinity CTL epitopes, it is critical to assess the impact of these changes on the conformation and flexibility of the modified epitope when bound to the MHCI, as well as on the specificity of the immune response elicited by the modified epitope.
Starting about a decade ago, techniques were developed that enabled the analysis of peptides bound to MHCIs (Falk et al., 1991, Nature 351, 290-296; Jardetsky et al., 1991, Nature 353, 326-329). Such techniques have been used extensively to characterize the peptides that bind different MHCI alleles. These studies have led to a general rule that each class I-bound peptide consists of between 8 and 11 amino acid residues. Studies have further shown that those peptides capable of forming a complex with a particular MHCI allele have in common the presence of 2 or 3 largely invariant residues at specific positions in the peptide. X-ray crystallographic studies have provided structural insight into the basis for these conserved residues. These studies have shown that each MHCI folds to form a groove, and it is within this groove that the peptide is displayed. The groove has a specific topology associated with each allele, and is characterized by the presence of several depressions or pockets along its length. Crystallographic studies have shown that the side chains of the conserved peptide residues (the “anchor residues”) extend into these pockets, and that the interaction between the MHCI and peptide within these pockets supplies a significant portion of the energy of binding. As such, these largely invariant, allele-specific residues are referred to as “primary anchor residues.” Interactions between side chains of amino acids near the termini of the peptide are also important for binding of the peptide to the MHCI, and they are observed in the crystallographic studies to bury into the ends of the MHCI binding groove. Much of the range in binding affinities among peptides that bind to a given MHCI allele is due to sequence variations in this array of primary anchor residues. However, it is also clear that other residues outside the primary anchor positions can contribute to MHCI binding (Kondo et al., 1995, J. Immunol. 155, 4307-4312; Schonbach et al., 1995, J. Immunol. 154, 5951-5958), and these other residues are referred to as “secondary anchor residues”. Secondary anchor residues may also contribute to the range in binding affinities among peptides binding to a particular MHCI.
A number of investigators have successfully enhanced the immunogenicity of both viral and tumor peptide epitopes by increasing the affinity of the peptide for the MHCI. Typically, this has been achieved by replacing sub-optimal residues found at the primary anchor positions with an optimal natural residue. In mice, an anchor-modified epitope from mutant Ras binds more effectively to the H-2Kd allele, induces enhanced cytolytic activity in vitro, and elicits a greater T-cell response in vivo than the unmodified parent mutant Ras peptide (Bristol et al., 1998, J. Immunol. 160, 2433-2441). Similar results were also observed in mice transgenic for human HLA-A2. In these experiments, an anchor-modified epitope from HIV reverse transcriptase was shown to be more effective at inducing CTL reactive with the reverse transcriptase than the parent peptide, and afforded greater protection in vivo against a challenge with Vaccinia virus expressing the HIV-1 reverse transcriptase (Okazaki et al., 2003, J. Immunol. 171, 2548-2555). In addition, anchor-modified epitopes from the melanoma antigen gp100 were shown to bind with higher affinity to HLA-A*0201 than did the parent epitopes, and induced melanoma-reactive CTL in peripheral blood leukocytes (PBL) isolated from 7 of 7 HLA-A2+ melanoma patients. In contrast, the unmodified parent epitopes induced melanoma-reactive CTL in PBL from only 2 of the 7 patients (Parkhurst et al., 1996, J. Immunol. 157, 2539-2548).
Alternative approaches to enhancing the binding affinity of CTL epitopes to MHCI include amino acid substitutions at positions other than the primary anchor residues. Systematic analyses of the role of secondary anchor residues in the binding of CTL epitopes to the MHCI have been performed, revealing that amino acid residues at positions other than the primary anchor residues can strongly influence the binding of the epitope (Deres et al., 1993, Cell. Immunol. 151, 158-167; Kondo et al., 1995, J. Immunol. 155, 4307-4312). In addition, studies examining the immunogenicity of two epitopes from ovalbumin indicate that amino acid residues at sites other than the primary anchor sites can significantly influence binding and presentation of an epitope (Chen et al., 1994, J. Exp. Med. 180, 1471-1483). A more general strategy has also been proposed to enhance the binding and immunogenicity of epitopes that bind with low affinity to HLA-A2.1. Studies with multiple low-affinity HLA-A2.1 epitopes show that the introduction of tyrosine at the first position can enhance the binding affinity and immunogenicity of the modified epitope (Pogue et al., 1995, Proc. Natl. Acad. Sci. USA 92, 8166-8170; Toudot et al., 2000, Eur. J. Immunol. 30, 3411-3421; Patent Application Publication US 2004/0072240 A1 by Kosmatopoulos et al.).
Non-natural amino acids have also been used with some success to replace natural amino acids and increase the immunogenicity of CTL epitopes. Using an antigenic peptide from the influenza virus matrix protein, a retro-inverso amide bond substitution was shown to enhance binding of the modified peptide to HLA-A2; however, this retro-inverso modification was only effective when inserted between Position 1 (P1) and Position 2 (P2) (as numbered from the amino terminus of the peptide), and resulted in a significant reduction in binding when used between P2 and P3, or P3 and P4, or P5 and P6, or P8 and P9 (Guichard et al., 1996, J. Med. Chem. 39, 2030-2039).
An alternative approach utilized an epitope from Melan-A (GenBank No. U06654). Melan-A is unrelated to any known gene and is encoded by an approximately 18-kilobase gene comprised of 5 exons (Coulie et al., 1994, J. Exp. Med. 180, 35-42). Similar to other melanoma-associated antigens, Melan-A is expressed in most melanoma tumor samples and normal melanocytes. HLA-A2 restricted melanoma-specific CTLs from nine of ten patients recognize primarily the MART-1 epitope, i.e., amino acid residues 27-35 (AAGIGILTV; SEQ ID NO:1) of Melan-A. Kawakami, Y., et al., J. Exp. Med. 180, 347-352 (1994). The MART-1 epitope of Melan-A has been studied extensively and has been used as a model system for investigating cellular immune responses in humans. Valmori et al., J. Immunol. 160, 1750-1758 (1998). The alternative approach involved the substitution of a j3-amino acid at Position 4 (P4), resulting in enhanced binding of the peptide to HLA-A2 and increased stability of the peptide-MHCI complex. However, only one out of three MART-1 specific CTL clones isolated from HLA-A2 restricted human tumor infiltrating lymphocytes (TIL) recognized and lysed target cells pulsed with the modified peptide. Guichard et al., J. Med. Chem. 43:3803-3808 (2000).
In an attempt to reduce biodegradation and enhance the immunogenicity of CTL epitopes, 36 peptide derivatives of the MART-1 epitope were designed based on knowledge of the degradation pathway in human serum. Non-natural amino acids were substituted at positions 1, 2, 8, 9 and 10 of the 10 amino acid MART-1 epitope individually and in combination and these modified MART-1 epitopes were then tested for their resistance to proteolysis and antigenic properties (Blanchet et al., 2001, J. Immunol. 167, 5852-5861). Eight of the modified epitopes were found to have enhanced stability against degradation when incubated in human serum, and three of these analogs were shown to be more potent than the parental epitope in stimulating in vitro MART-1 specific CTL responses in PBMC from normal donors. Non-natural amino acids have also been used in studies aimed at enhancing the immunogenicity of a poorly immunogenic epitope from murine p53. By replacing cysteine with aminobutyric acid at positions 4 and 8 and methionine with norleucine at positions 3 and 9 in the epitope, the modified peptides bound with higher affinity to MHCI and appeared to be more potent immunogens. Baratin et al., J. Peptide Sci. 8:327-334 (2002).
With a strong scientific rationale, highly promising preclinical results and the frequently observed induction of target-specific immune responses in treated patients, the interest in and efforts directed at the development of effective cancer vaccines continues to increase. A recent study described 645 clinical trials related to cancer vaccines and reported that new cancer vaccine trials had shown a steady increase since 2001 reaching more than 60 new trials each year (Cao, X., et al., 2008, Immunome Research 4, 1-11). However, the results from these clinical trials have been much less encouraging with only rare occurrences of objective tumor regression despite the detection in some patients of robust target-specific immune responses (Mocellin, S., et al., 2009, Curr. Med. Chem. 16, 4779-4796). Investigators have identified a number of factors that likely contribute to the poor clinical results including the selection of suboptimal targets, the presence of mutations in tumor cells that can promote tumor escape and immunosuppressive factors including cells, proteins and chemicals present in the tumor environment. It has been proposed that a successful cancer vaccine will need to address these and other factors by incorporating at least one and most likely multiple optimized epitopes, enhanced delivery systems, adjuvants and co-factors to activate cytotoxic and helper T-cells as well as antigen presenting cells, and strategies to inhibit the immunosuppressive network (Kanodia, S. and Kast, W. M., 2008, Expert Rev. Vaccines 7, 1533-1545).
There is a continuing need in the art for improved methods of generating a T-cell immune response.